Symeon Charalabides

Crime rates have been on the increase in different parts of the world and every country has been looking for more effective and efficient ways of tackling crimes in their societies. The advancement in information technology has improved crime data collection and the “Internet of Things” has given rise to an explosion of data that is easily available [1]. This has led to an enormous amount of data that the law enforcement agencies must go through in order to gain insight on the trends and… Show more

Crime rates have been on the increase in different parts of the world and every country has been looking for more effective and efficient ways of tackling crimes in their societies. The advancement in information technology has improved crime data collection and the “Internet of Things” has given rise to an explosion of data that is easily available [1]. This has led to an enormous amount of data that the law enforcement agencies must go through in order to gain insight on the trends and patterns of crime occurrences. However, security is one of the major threats in the modern world and without security in New York City; it cannot thrive economically and socially. For this reason, policy makers and law enforcement agencies need greater understanding of the New York Police Department (NYPD) crime dataset to help them fight the root cause of crime, effectively control crime rate and protect their citizens. This project is aimed at comparing the classification accuracy of six different types of machine learning algorithms in order to select the best model for crime prediction. The algorithms are Support Vector Machine (SVM), Random Forest (RF), Classification and Regression Analysis (CART), Linear Discriminant Analysis (LDA), Naive Bayes (NB), and K-Nearest Neighbour (KNN). Based on the results, RF and CART achieved 100% accuracy and kappa value of 1, both in the training and testing of the models. SVM achieved accuracy of 99.98%. LDA and NB achieved classification accuracies of 99.76% and 91.76% respectively, while KNN has the lowest accuracy value at 55.74%. Show less

Breast cancer is one of the most common types of cancer in Ireland and worldwide. Any effort that helps to obtain an early diagnosis or preventing any cancer cell growth is helpful. This idea inspires this project. Using data from the Breast Cancer Wisconsin's Data Set (UCI Machine Learning), we use machine learning techniques to predict the existence of any cancer cells. New technologies such as data storage using the Hadoop system on AWS (Amazon), clustering and several linear and non-linear… Show more

Breast cancer is one of the most common types of cancer in Ireland and worldwide. Any effort that helps to obtain an early diagnosis or preventing any cancer cell growth is helpful. This idea inspires this project. Using data from the Breast Cancer Wisconsin's Data Set (UCI Machine Learning), we use machine learning techniques to predict the existence of any cancer cells. New technologies such as data storage using the Hadoop system on AWS (Amazon), clustering and several linear and non-linear prediction methods are used to diagnose the condition of the cell (and the patient). The different results are compared based on accuracy performance, confusion matrix and area under the Receiver Operating Characteristics (ROC) curve. A classification error means sending a patient home who could potentially have cancer. Therefore, minimizing classification errors is vital in this approach. The goal of this project is to find one or more methods to solve the problem. The best models are chosen using performance metrics such as the area under the ROC curve and the area under the Precision-Recall curve and prediction accuracy. Show less

The EGRET gamma-ray telescope has left a legacy of unidentified astronomical sources. Most likely, many of the galactic plane sources will be rotation-powered pulsars. Firm identification has been difficult, given the instrument's poor spatial resolution. The problem is exacerbated by the energy dependant Point Spread Function (PSF) and low numbers of source counts. The main method of identifying sources to-date has been a maximum likelihood method. We have taken a different approach, namely… Show more

The EGRET gamma-ray telescope has left a legacy of unidentified astronomical sources. Most likely, many of the galactic plane sources will be rotation-powered pulsars. Firm identification has been difficult, given the instrument's poor spatial resolution. The problem is exacerbated by the energy dependant Point Spread Function (PSF) and low numbers of source counts. The main method of identifying sources to-date has been a maximum likelihood method. We have taken a different approach, namely that of regularized deconvolution with a spatially invariant PSF, which is used in optical astronomy and medical X-ray imaging. This technique revealed that wavelet denoising of residuals produced smooth, relatively artefact-free images with improved spatial location. Our source location using standard centroiding produced an improvement in relative spatial location, ranging from 10:1 to 2:1 proportional to source strength. Wavelet deconvolution simultaneously achieves background smoothing, while improving sharpness of the resolved objects. The photon-sparse nature of these images makes them an ideal test bed for such techniques. Although deconvolution does not ordinarily conserve flux, in this instance the flux determination is unaffected in all but the most crowded regions. Finally, we show that the energy dependent PSF can be used to identify objects with a restricted range of energy spectra. Show less